Poke vs Hermes Agent (2026)

Poke

by The Interaction Company of California

Hermes Agent

by Nous Research

Pros

• ✓First and only AI agent approved for Apple Messages for Business, giving Poke native access to iMessage: one billion iPhone users can interact with a personal AI agent inside their default messaging app without downloading anything, a distribution advantage no other personal AI agent has achieved.
• ✓Model-agnostic architecture selects the best AI model per task from OpenAI, Anthropic, Google, and open source providers: unlike ChatGPT (locked to OpenAI), Meta AI (locked to Meta), and Gemini (locked to Google), Poke is not bound to a single model provider and can route to whichever model performs best for each specific request.
• ✓Poke Recipes and Kitchen system enables open-ended extensibility via MCP servers: developers can build custom integrations using the official GitHub template (InteractionCo/mcp-server-template), and users share community-built recipes across productivity, health, developer tools, finance, and travel categories, making the integration surface grow without waiting for official connectors.

Limitations

• ⚠No enterprise governance, admin controls, SSO, or audit logging: Poke is a consumer personal assistant without the compliance and security infrastructure that Microsoft Scout (governed Entra identity, Purview, Defender XDR) and Gemini Spark (Google enterprise controls) provide for regulated organizations.
• ⚠Pricing is negotiable via the onboarding "bouncer" chat, which some users find aggressive and off-putting: reported prices range from $3 to $30/month depending on the negotiation, creating inconsistency that the published $19/month Pro and $199/month Ultra tiers do not fully resolve.
• ⚠No G2, Capterra, or independent reviews as of Q2 2026, and investor funding amount has not been disclosed: while notable angels (Stripe founders, DeepMind, OpenAI, Cognition) signal insider confidence, the absence of public adoption metrics beyond 100M messages makes it difficult to assess real-world reliability at scale.

Pros

• ✓Self-improving learning loop with no manual upkeep: after each complex task the agent automatically creates and refines SKILL.md files so it never forgets how to solve recurring problems, and the 647-skill community ecosystem means most common workflows have a starting point without any user configuration
• ✓Runs on infrastructure you control with zero telemetry, zero tracking, and zero data leaving your machine: a meaningful security and privacy advantage over SaaS agents for teams handling sensitive data, proprietary research, or regulated information
• ✓Most widely adopted open-source AI agent with {{github_stars}} GitHub stars: MCP compatible with full cross-session memory across 20 plus platforms, with a desktop app and web portal now available alongside the original CLI

Limitations

• ⚠CLI-first setup with moderate technical requirements: deployment needs a server or VPS, familiarity with a terminal, and an LLM API key; there is no hosted SaaS version or graphical setup wizard, which limits accessibility for non-technical users who cannot configure a Linux environment
• ⚠No built-in cost controls on LLM API usage: the agent runs autonomously and will continue making API calls during scheduled tasks and multi-step workflows, which can generate unexpected token costs without careful monitoring of usage and provider spending limits
• ⚠Memory system uses character-limited files injected as a frozen snapshot at session start rather than a vector database: keeps the system lightweight and predictable but means very large or rapidly growing memory contexts require manual curation to stay within limits